On Mon, 17 Oct 2022 13:25:13 +0000 Tim Chase via Remind-fans <[email protected]> wrote:
> I just compared the clocks on my OpenBSD machines using > > $ TZ=America/Toronto date > Mon Oct 17 09:15:24 EDT 2022 Yes, but that's not the test we need. We need to know what time it thinks it is at 2022-11-06T01:59 in Toronto. Daylight Saving Time in Toronto ends at 2022-11-06T02:00 local time. However, there's an ambiguity and it's possible OpenBSD's C library and glibg handle it differently. There are actually *two different* times in Toronto at 2022-11-06T01:59: One before daylight saving time ends, which corresponds to the time *before* DST ends, which means 2022-11-06T05:59 UTC. The second 2022-11-06T01:59 corresponds to the time *after* DST ends, which is 2022-11-06T06:59. glibc appears to use the first interpretation while OpenBSD's libc appears to use the second interpretation. I guess I should probably change the tests to avoid that ambiguous time window. > I had similar complaints from `make test` on my FreeBSD daily-driver > with a LOT of timezone hiccups being off by 1hr. I've attached my > tests/test.out in case that helps track down what's going on. See above... I bet that's it. Regards, Dianne. _______________________________________________ Remind-fans mailing list [email protected] https://dianne.skoll.ca/mailman/listinfo/remind-fans Remind is at https://dianne.skoll.ca/projects/remind/
